Food webs are the simplest type of biological networks that describes the trophic interaction between species. They possess non-random structural complexities and how to statistically infer them remains as an open question. In this seminar, I present some models that attempt to explain real food webs by using simple assembly rules. As we will see, those models do not sufficiently capture the complexity of real food webs, and therefore statistical inference using those models can be problematic. To this end, I propose from ecological intuition a model for inferring the structural properties of food webs. This approach generates samples of hypothetical food webs by bootstrapping from real food web data. Subsequently, those bootstrap samples are then used to generate the sampling distribution of several food web statistics, which can be then used for statistical inference.
Wei-chung Liu is currently an associate research fellow in Institute of Statistical Science, Academia Sinica. He is an ecologist and has some interests in the theoretical aspects of biology and sociology.